Chronic disease patient information management system and method based on neural convolution algorithm
The generation of body state coefficients and matching rehabilitation plans through neural convolution algorithms solves the problem that existing systems are difficult to manage health data of chronic disease patients, realizes effective health management and rehabilitation assistance for chronic disease patients, and improves information management and rehabilitation efficiency .
Patent Information
- Application Number
- CN202411198348.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-29
AI Technical Summary
The existing patient information management system is difficult to effectively manage the health data of chronic patients and cannot achieve a rehabilitation auxiliary role, especially when the number of visits is frequent and health data is updated frequently.
Using a chronic disease patient information management system based on neural convolution algorithm, the patient is marked as non-healthy or sub-healthy by generating body state coefficients, setting a health range, and matching the rehabilitation plan according to health indicators and key factors, and optimizing the rehabilitation plan to Accelerate patient recovery.
It improves the pertinence and efficiency of information management, can distinguish and manage different patients, reduce management difficulty, assist or accelerate the rehabilitation process, deal with health abnormalities in advance, and optimize rehabilitation plans to improve patients' health status.
Smart Images

Figure CN118711739B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of patient information management, and specifically to a chronic disease patient information management system and method based on a neural convolution algorithm. Background Art
[0002] Chronic diseases are a type of chronic disease with special characteristics that have a profound impact on people's health and quality of life. The full name is chronic non-communicable diseases. It does not refer to a specific disease, but is a general term for a type of disease with hidden onset, long course and prolonged illness, lack of definite evidence of infectious biological causes, complex causes and some diseases that have not yet been fully confirmed. The characteristics of this type of disease mainly include relatively stable condition, mild symptoms, long course of illness, easy recurrence and difficulty in cure. Since chronic diseases are usually closely related to unhealthy lifestyles, such as smoking, excessive drinking, high-salt and high-fat diets, and too little exercise, actively changing bad lifestyles and eating habits is the key to preventing and managing chronic diseases. In addition, regular physical examinations, early detection, and early treatment are also important means to control the development of chronic diseases.
[0003] Patient information management is the process by which medical institutions use electronic means to systematically collect, organize, store, transmit, utilize and share patients' basic information, medical records and diagnostic results. It aims to improve the efficiency of medical services, protect the rights and interests of patients, and promote the security and sharing of medical information. It is an indispensable and important task for medical institutions. It runs through the entire process of patient diagnosis and treatment, and can improve the quality of medical services and protect the rights and interests of patients.
[0004] In the Chinese invention patent with application number 202410104025.X, a patient information management method and system based on smart medical care are provided, each target patient information is classified and processed, and the classification result corresponding to each target patient information is obtained, and then the pending items are managed and processed according to each target patient information and each classification result, so as to obtain the management and processing results of the pending items. First, the appropriate original patient information is extracted as the target patient information, and the data that does not meet the evaluation standards is deleted to a certain extent. Then, on the premise of analyzing that the pending items are target items, all target patient information is classified and processed, so that the corresponding classification results can contain more detailed element information of the pending items. In this way, the pending items are managed and processed according to each target patient information and each classification result, so that the management and processing results of the pending items can be analyzed more accurately, and the reliability and efficiency of managing the item elements can be improved.
[0005] Combined with the above application and the contents of the prior art:
[0006] Considering that the recovery period of patients with chronic diseases is long and they may need to see a doctor frequently, patient information may need to be used frequently. Therefore, the relevant information of patients with chronic diseases needs to be managed in a targeted manner. Existing patient information management methods are usually targeted at information security. After medical staff complete data collection, they usually set access rights or data encryption for users to prevent the loss of patient information and avoid serious security consequences.
[0007] However, when chronic disease patients visit the doctor more frequently, the frequency of communication between doctors and patients is high, and the frequency of updating of patient health data is high, the difficulty of patient information management and the need for in-depth management are further increased. However, after collecting and obtaining relevant information from patients, the existing information management system can only be used for information security management, but it is difficult to achieve health management of chronic disease patients and cannot play an auxiliary role in the rehabilitation of chronic disease patients.
[0008] To this end, the present invention provides a chronic disease patient information management system and method based on a neural convolution algorithm. Summary of the invention
[0009] 1. Technical issues to be resolved
[0010] In view of the deficiencies of the prior art, the present invention provides a chronic disease patient information management system and method based on a neural convolution algorithm, which uses the patient health index data to generate a physical state coefficient, marks the corresponding patient as an unhealthy patient or a sub-healthy patient, sets a health range for the sub-healthy patient, and uses the health index and the corresponding key factors as input to match the corresponding rehabilitation plan for the unhealthy patient; if there are abnormal indicators in the obtained predicted values, the aggregation of the abnormal indicators is generated according to the distribution state of the abnormal indicators, and the patients are re-differentiated according to the aggregation, and the rehabilitation plan for the unhealthy patients is optimized. Taking the optimized rehabilitation plan as a reference, the patient's rehabilitation is accelerated when the patient's health is managed, thereby solving the technical problems raised in the background technology.
[0011] (II) Technical solution
[0012] To achieve the above objectives, the present invention is implemented through the following technical solutions:
[0013] The information management method for chronic disease patients based on neural convolution algorithm includes: after collecting the patient information management requirements, generating the demand complexity from the patient's information management requirements data set If the complexity of the requirement If the complexity threshold is exceeded, a data collection instruction is issued to the outside; among them, the communication frequency and data update frequency Perform linear normalization and then follow the following method:
[0014] ;
[0015] Weight coefficient, , ;
[0016] After collecting and entering patient-related data, construct a patient information data set, and analyze and obtain the diversity of patient information in the patient information data set , according to diversity Matching a corresponding data encryption scheme from a data encryption scheme library to encrypt the stored patient information;
[0017] According to the data in the patient information data set, the patients are divided into different patient categories, and according to the patient-related data and disease characteristics, the trained decision tree algorithm is used to identify the key factors affecting the patient's health, and the identified key factors are output;
[0018] Generate physical status coefficients from patient health indicator data , according to the health status coefficient Mark the corresponding patients as unhealthy patients or sub-healthy patients, set the health range for sub-healthy patients, and use the health indicators and corresponding key factors as input to match the corresponding rehabilitation programs for unhealthy patients;
[0019] After collecting the patient's health indicator data, prediction is performed. If there are abnormal indicators in the obtained prediction values, the clustering of the abnormal indicators is generated according to the distribution state of the abnormal indicators. The patients are re-differentiated according to the clustering, and the rehabilitation plan for unhealthy patients is optimized.
[0020] Furthermore, when managing the information of chronic disease patients, the frequency of communication between doctors and patients, the frequency of updating of patients' health data, and the change status data of patients' health status are counted within the preset patient information management cycle, and the information management demand data set is summarized to generate the demand complexity. .
[0021] Furthermore, after receiving the data collection instruction, the patient's basic information, medical record form, examination result form and follow-up record form are collected to build a patient information data set, and the data access rights are set after the patient information data set is stored; several data encryption schemes with different cracking difficulties are obtained, and a data encryption scheme library is pre-built.
[0022] Furthermore, the diversity of patient information in the patient information data set is calculated , where the similarity between the two patient information Perform linear normalization and map the corresponding data values to the interval In the following way:
[0023] ;
[0024] in, is the number of patients, It is i The patient and j The similarity between the patient information of each patient, is the average similarity; weight coefficient: , ,and .
[0025] Furthermore, the data in the patient information data set is preprocessed to obtain the preprocessed patient information to label the patients; feature engineering is used to extract features of the patient information in each patient class to obtain corresponding disease characteristics; and relevant data of each patient class is collected.
[0026] Furthermore, the health indicator data of the patient in the current stage is collected and summarized to generate the patient's health indicator set; the patient's physical condition coefficient is generated from the health indicator set , if the patient's health status coefficient If the patient's blood pressure is lower than the health threshold, the corresponding patient is marked as an unhealthy patient, otherwise, he / she is marked as a sub-healthy patient.
[0027] Furthermore, blood glucose concentration and blood pressure Perform linear normalization and generate the body state coefficient as follows :
[0028] ;
[0029] Weight coefficient: , ; is the number of collection points, For the i Blood pressure at each collection node, is the mean blood pressure, is its qualified standard value; For the i The blood glucose concentration at each collection node is is the mean blood glucose concentration, is the qualified standard value.
[0030] Furthermore, a healthy range is set for sub-healthy patients The way is as follows:
[0031] ;
[0032] in, , is the number of predicted values of health status coefficient, is the mean of the predicted values of the health status coefficient; For the i The predicted value of the health status coefficient on the prediction node.
[0033] Furthermore, the current sets of health indicator data are used as input, and the trained neural convolutional network is used to predict the health trend of sub-healthy patients and obtain the corresponding predicted values; the health range is set for sub-healthy patients based on the obtained predicted values. ; If the predicted value of the health status coefficient of the sub-healthy patient is not within the healthy range Within, a prompt instruction is issued to the outside.
[0034] Furthermore, the acquired rehabilitation plans are summarized to construct a rehabilitation plan library, and the health indicators and corresponding key factors of non-healthy patients are used as input. A collaborative filtering algorithm is used to match corresponding rehabilitation plans for non-healthy patients from the rehabilitation plan library, and the rehabilitation plans are sent to the non-healthy patients.
[0035] Furthermore, after a preset vital sign data collection cycle, wearable devices are used to monitor the patient's health indicator data in real time, and a patient health data set is generated by aggregation; a time series prediction model is used to predict each health indicator data, and the corresponding indicator prediction value is obtained; the indicator prediction value is used as input, and the trained abnormal indicator recognition model is used for recognition, the identified abnormal indicator is output, and the time node when the abnormal indicator is generated is determined.
[0036] Furthermore, combined with the distribution state of abnormal indicators in the current vital sign data collection period, the aggregation degree of abnormal indicators is generated according to the following method: :
[0037] ;
[0038] Weight coefficient: , ,and ; is the number of abnormal indicators, It is i The abnormal indicators and j The time interval between abnormal indicators, is the average value of the time interval, is the acceptable value for the time interval;
[0039] If the concentration If the concentration threshold is exceeded and the patient is sub-healthy, the sub-healthy patient will be marked as an unhealthy patient. If the patient is an unhealthy patient, a program optimization instruction will be issued to the outside.
[0040] Furthermore, after receiving the program optimization instruction, the change status data of the health indicators of non-healthy patients after executing the rehabilitation program and the feedback on the rehabilitation program are collected, and the reduction of the density of abnormal indicators is used as the optimization goal. The rehabilitation program of non-healthy patients is optimized using the trained multi-objective optimization algorithm to obtain the optimized rehabilitation program.
[0041] The chronic disease patient information management system based on the neural convolution algorithm includes a demand analysis unit, which collects patient information management requirements and generates demand complexity from the patient information management requirement data set. If the complexity of the requirement If the complexity threshold is exceeded, a data collection instruction is issued to the outside;
[0042] Diversity analysis unit, collects and enters patient-related data to build a patient information data set, and analyzes and obtains the diversity of patient information in the patient information data set , according to diversity Matching a corresponding data encryption scheme from a data encryption scheme library to encrypt the stored patient information;
[0043] The patient classification unit classifies patients into different patient categories based on the data in the patient information data set, and uses the trained decision tree algorithm to identify key factors affecting the health of patients based on the patient-related data and disease characteristics, and outputs the identified key factors;
[0044] Health status assessment unit, which generates body status coefficients from patient health index data , according to the health status coefficient Mark the corresponding patients as unhealthy patients or sub-healthy patients, set the health range for sub-healthy patients, and use the health indicators and corresponding key factors as input to match the corresponding rehabilitation programs for unhealthy patients;
[0045] The prediction optimization unit collects the patient's health indicator data and then makes a prediction. If there are abnormal indicators in the obtained prediction values, the clustering of the abnormal indicators is generated according to the distribution state of the abnormal indicators. The patients are re-differentiated according to the clustering, and the rehabilitation plan for unhealthy patients is optimized.
[0046] (III) Beneficial effects
[0047] The present invention provides a chronic disease patient information management system and method based on a neural convolution algorithm, which has the following beneficial effects:
[0048] 1. Based on the complexity of requirements Evaluate the difficulty and complexity of current patient information management, and adopt targeted management methods to improve the pertinence of information management when managing patient information; Assessing the similarity between different patient information can help provide targeted protection.
[0049] 2. Based on the diversity of patient information Choosing a data encryption solution can provide targeted protection for patient information and reduce the risk of information leakage; after identifying the patient's relevant data, determine the key factors that affect the patient, and make targeted adjustments to the key factors when managing patient information, so that the adjusted key factors are more compatible with the patient's physical health indicators, avoiding further negative impacts on the patient's health.
[0050] 3. By differentiating patients into different categories, targeted management of different patients can be achieved, and the rehabilitation efficiency of patients can be improved by improving the targeted management; on the basis of collecting the health indicator data of patients, patients can be divided into sub-healthy patients and unhealthy patients, so as to guide managers or medical staff to carry out differentiated management when managing patient information and reduce the difficulty of management; on the basis of predicting the health trends of sub-healthy patients, by setting a healthy range for them, it is helpful to establish a rehabilitation plan for sub-healthy patients to assist or accelerate their rehabilitation.
[0051] 4. By matching and recommending pre-built rehabilitation plans for non-healthy patients in a targeted manner, and using rehabilitation plans as a reference when managing patient information, the difficulty of information management and formulation of rehabilitation plans can be reduced. At the same time, it can also accelerate the patient's rehabilitation progress and improve the efficiency of information management.
[0052] 5. Through wearable devices, the patient's current health indicators can be monitored. The acquired health indicator data can be used to judge the patient's current health status. The subsequent data can be predicted to grasp the patient's health change trend and deal with possible abnormal situations in advance.
[0053] 6. The trained abnormal indicator recognition model is used for identification. By determining whether there are abnormal indicators, it can be judged that the patient's current health indicators are abnormal and can be treated in time; by generating the aggregation degree of abnormal indicators , to determine the abnormality of the patient's current health indicators, and to re-divide and adjust the patients accordingly, so as to maintain targeted management of the patients; for non-healthy patients, if the abnormality of their health indicators is still high, by optimizing the rehabilitation plan and using the optimized rehabilitation plan as a reference, the patient's recovery can be accelerated when managing their health. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] Figure 1 This is a flow chart of the method for managing information of chronic disease patients of the present invention;
[0055] Figure 2 This is a schematic diagram of the structure of the chronic disease patient information management system of the present invention. DETAILED DESCRIPTION
[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0057] See also Figure 1 The present invention provides a method for managing information of chronic disease patients based on a neural convolution algorithm, comprising:
[0058] Step 1: After collecting patient information management requirements, generate the demand complexity from the patient information management requirements data set If the complexity of the requirement If the complexity threshold is exceeded, a data collection instruction is issued to the outside;
[0059] The step 1 includes the following contents:
[0060] Step 101: When managing the information of chronic disease patients, the frequency of communication between doctors and patients, the frequency of updating of patient health data, and the change status data of patient health status are counted within a preset patient information management cycle, and a collection of information management demand data is generated;
[0061] Step 102: Generate demand complexity from the patient's information management demand data set , where the communication frequency and data update frequency Perform linear normalization and then follow the following method:
[0062] ;
[0063] Weight coefficient, , ;The weight coefficient can be obtained by referring to the hierarchical analysis method;
[0064] Based on historical data and the needs for patient information management, the complexity threshold is pre-set; if the complexity of the requirements is If the complexity threshold is exceeded, it means that the various requirements that need to be managed in the current stage are highly complex, and the need for targeted management is also high. At this time, a data collection instruction is issued to the outside.
[0065] When using, combine the contents in steps 101 and 102:
[0066] When it is necessary to manage various information of chronic disease patients, the usage status of patient information is collected, and the demand complexity is generated based on the collection results. , according to the complexity of the requirements Evaluate and describe the current difficulty and complexity of patient information management, so that targeted management methods can be adopted when managing patient information to improve the targeted nature of information management.
[0067] However, when chronic disease patients visit the doctor more frequently, the frequency of communication between doctors and patients is high, and the frequency of updating of patient health data is high, the difficulty of patient information management and the need for in-depth management are further increased. However, after collecting and obtaining relevant information from patients, the existing information management system can only be used for information security management, but it is difficult to achieve health management of chronic disease patients and cannot play an auxiliary role in the rehabilitation of chronic disease patients.
[0068] Step 2: After collecting and entering patient-related data, construct a patient information data set and analyze the diversity of patient information in the patient information data set , according to diversity Matching a corresponding data encryption scheme from a data encryption scheme library to encrypt the stored patient information;
[0069] The step 2 includes the following contents:
[0070] Step 201, after receiving the data collection instruction, collect the patient's basic information: name, gender, age, contact information, address, etc.; medical record form: disease type, diagnosis date, treatment plan, past medical history, etc.; examination result form: including the results and dates of various examinations (such as blood sugar, blood pressure, blood lipids, etc.); follow-up record form: record the date of each follow-up, doctor's advice, patient feedback, symptom changes, etc.;
[0071] The above collected data is automatically input using OCR technology to complete the construction of the patient information data set. After the patient information data set is stored, data access permissions are set to ensure that only authorized personnel can access patient information;
[0072] When using, set the corresponding data access permissions after completing data collection and storage to avoid data leakage and loss;
[0073] Step 202: Calculate the diversity of patient information in the patient information data set , where the similarity between the two patient information Perform linear normalization and map the corresponding data values to the interval In the following way:
[0074] ;
[0075] in, is the number of patients, It is i The patient and j The similarity between the patient information of each patient, is the average similarity; weight coefficient: , ,and ;
[0076] When used, by calculating the diversity of patient information , according to diversity Evaluate the similarity between different patient information. If the similarity between patient information is lower and the complexity is higher, the importance of the corresponding data may be higher, so it needs to be protected in a targeted manner;
[0077] Step 203: by online retrieval or offline formulation, a number of data encryption schemes with different cracking difficulties are collected and a data encryption scheme library is pre-built; according to the positive correlation between the cracking difficulty of the data encryption scheme and the diversity of the patient information, a data encryption scheme of a corresponding level is matched from the data encryption scheme library, and the data encryption scheme is executed to encrypt the stored patient information;
[0078] When using, combine the contents in steps 201 to 203:
[0079] When patient information needs to be protected, Choosing a data encryption solution can provide targeted protection for patient information and reduce the risk of information leakage.
[0080] Step 3: Classify patients into different patient categories based on the data in the patient information data set, and use the trained decision tree algorithm to identify key factors affecting the health of patients based on the patient-related data and symptom characteristics, and output the identified key factors;
[0081] The step three includes the following contents:
[0082] Step 301: preprocess the data in the patient information data set, such as data cleaning, data conversion, data integration and data visualization, take the preprocessed patient information to label the patients, use the cluster analysis algorithm to classify the patients, and distinguish the patients into different patient classes;
[0083] When in use, after obtaining the patient's identity information, targeted management of different patients is achieved by differentiating the patients into different categories, thereby improving the patient's rehabilitation efficiency by improving the targeted management.
[0084] Step 302: extract features from the patient information in each patient class using feature engineering to obtain corresponding disease features; collect relevant data of each patient class, such as daily routine and eating habits, and various health index data, and use the patient's disease features and related data as input to use the trained decision tree algorithm to identify key factors that affect the patient's health, and output the identified key factors;
[0085] When used, combine the contents in steps 301 and 302:
[0086] After identifying the patient's relevant data, determine the key factors that affect the patient. On this basis, make targeted adjustments to the key factors when managing patient information so that the adjusted key factors are more compatible with the patient's physical health indicators, thereby avoiding further negative impact on the patient's health.
[0087] Step 4: Generate physical condition coefficients from patient health index data , according to the health status coefficient Mark the corresponding patients as unhealthy patients or sub-healthy patients, set the health range for sub-healthy patients, and use the health indicators and corresponding key factors as input to match the corresponding rehabilitation programs for unhealthy patients;
[0088] The step 4 includes the following contents:
[0089] Step 401: collect and obtain the health index data of the patient at the current stage, such as weight, blood sugar, blood lipids and blood pressure, and generate a health index set of the patient after aggregation;
[0090] Generate the patient's physical condition coefficient from the health indicator set The patient's current physical condition is comprehensively judged by the physical condition coefficient, among which the blood glucose concentration and blood pressure Perform linear normalization and then follow the following method:
[0091] ;
[0092] Weight coefficient: , ; The weight coefficient can be obtained by referring to the hierarchical analysis method, is the number of collection points, For the i Blood pressure at each collection node, is the mean blood pressure, is its qualified standard value; For the i The blood glucose concentration at each collection node is is the mean blood glucose concentration, is its qualified standard value;
[0093] Based on historical data and patient health management expectations, health thresholds are pre-set; if the patient's health status coefficient If the patient is above the health threshold, the corresponding patient is marked as an unhealthy patient, otherwise, he is marked as a sub-healthy patient;
[0094] When in use, the patients are divided into sub-healthy patients and unhealthy patients based on the collection of their health indicator data, thereby guiding managers or medical staff to conduct differentiated management when managing patient information and reducing the difficulty of management.
[0095] Step 402: For sub-healthy patients, the current sets of health index data are used as input, and the trained neural convolution network is used to predict the health trend of the sub-healthy patients to obtain corresponding predicted values; and a health range is set for the sub-healthy patients based on the obtained predicted values. , as follows:
[0096] ;
[0097] in, , is the number of predicted values of health status coefficient, is the mean of the predicted values of the health status coefficient; For the i The predicted value of the health status coefficient on the prediction node;
[0098] If the predicted value of the health status coefficient of the sub-healthy patient is not within the healthy range Within, send prompt instructions to the outside;
[0099] When in use, based on the prediction of the health trends of sub-healthy patients, by setting a health range for them, it helps to establish a rehabilitation plan for sub-healthy patients, assist or accelerate their recovery.
[0100] Step 403: For non-healthy patients, several rehabilitation plans are collected in advance, such as personalized diet, exercise plan and work and rest plan, and the acquired rehabilitation plans are summarized to build a rehabilitation plan library. The health indicators and corresponding key factors of non-healthy patients are used as inputs, and a collaborative filtering algorithm is used to match the corresponding rehabilitation plan for the non-healthy patients from the rehabilitation plan library, and the rehabilitation plan is sent to the non-healthy patients;
[0101] When used, combine the contents in steps 401 to 403:
[0102] By specifically matching and recommending pre-built rehabilitation plans for non-healthy patients, and using the rehabilitation plans as a reference when managing patient information, the difficulty of information management and rehabilitation plan formulation can be reduced. At the same time, it can also accelerate the patient's rehabilitation progress and improve the efficiency of information management.
[0103] Step 5: After collecting the patient's health indicator data, make a prediction. If there are abnormal indicators in the obtained prediction values, generate the aggregation of abnormal indicators according to the distribution state of the abnormal indicators, re-differentiate the patients according to the aggregation, and optimize the rehabilitation plan for non-healthy patients;
[0104] The step five includes the following contents:
[0105] Step 501: After a preset vital sign data collection cycle, a wearable device, such as a smart watch, a blood glucose monitor, etc., is used to monitor the health index data of patients, i.e., sub-healthy patients and unhealthy patients, in real time, including heart rate, blood pressure, blood glucose, heart rate variability, blood oxygen saturation and body temperature, etc. After continuously acquiring several groups of health index data, a patient health data set is generated;
[0106] Step 502: Use the time series prediction model to predict various health indicator data and obtain the corresponding indicator prediction values; obtain an abnormal indicator recognition model through neural convolution network training, use the indicator prediction values as input, use the trained abnormal indicator recognition model to recognize, output the recognized abnormal indicators and determine the time node when the abnormal indicators are generated;
[0107] When in use, after setting rehabilitation goals or outputting rehabilitation plans, the wearable device can be used to monitor the various health indicators of the patient at the current stage, and the obtained health indicator data can be used to judge the patient's current health status;
[0108] At the same time, based on the current data and the corresponding historical data, the next data can be predicted to grasp the health change trend of the patient, and possible abnormal situations can be handled in advance when managing patient information;
[0109] Step 503: Based on the distribution state of abnormal indicators in the current vital sign data collection period, the aggregation degree of abnormal indicators is generated in the following manner: , the aggregation degree of abnormal indicators is evaluated by aggregation, the method is as follows:
[0110] ;
[0111] Weight coefficient: , ,and ;The weight coefficient value remains consistent with the previous value;
[0112] is the number of abnormal indicators, It is i The abnormal indicators and j The time interval between abnormal indicators, is the average value of the time interval, is the acceptable value for the time interval;
[0113] Pre-set aggregation thresholds based on the degree of management of patient health based on historical data;
[0114] If the concentration If the concentration threshold is exceeded, it means that the patient's physical health has not improved. If the patient is a sub-healthy patient, the sub-healthy patient will be marked as an unhealthy patient. If the patient is an unhealthy patient, it means that the effectiveness of the current rehabilitation program is insufficient. At this time, an instruction to optimize the program is issued to the outside.
[0115] When in use, based on the obtained prediction data, the trained abnormal indicator recognition model is used for identification. By determining whether there are abnormal indicators, it can be judged that the patient's current health indicators are abnormal, and by issuing reminders to the outside, it can be dealt with in time; further, by generating the aggregation degree of abnormal indicators , to determine the abnormality of the patient's current health indicators, verify the patient's current health level, and re-divide and adjust the patients based on this, so as to maintain targeted management of the patients.
[0116] Step 504: after receiving the program optimization instruction, collect the change state data of the health indicators of the non-healthy patients after executing the rehabilitation program and the feedback on the rehabilitation program, combine the above feedback data, take reducing the density of abnormal indicators as the optimization goal, use the trained multi-objective optimization algorithm to optimize the rehabilitation program of the non-healthy patients, and obtain the optimized rehabilitation program;
[0117] When using, combine the contents in steps 501 to 504:
[0118] For non-healthy patients, if their health indicators are still abnormal to a high degree, it means that the current implementation plan still has certain shortcomings. By optimizing the rehabilitation plan and using the optimized rehabilitation plan as a reference, the patient's recovery can be accelerated when managing the patient's health.
[0119] The analytic hierarchy process is an analytical method that combines qualitative and quantitative methods. It can decompose complex problems into multiple levels. By comparing the importance of factors at each level, it can help decision makers make decisions on complex problems and determine the final decision plan. In this process, the analytic hierarchy process can be used to determine the weight coefficients of these indicators. Through the analytic hierarchy process, decision makers can decompose complex decision-making problems into different levels and make decisions based on qualitative and quantitative analysis. This method can improve the accuracy and effectiveness of decision-making, and is particularly suitable for complex problems that are difficult to solve with quantitative methods.
[0120] See also Figure 2 The present invention provides a chronic disease patient information management system based on a neural convolution algorithm, including a demand analysis unit, which collects patient information management requirements and generates a demand complexity from the patient's information management requirement data set. If the complexity of the requirement If the complexity threshold is exceeded, a data collection instruction is issued to the outside;
[0121] Diversity analysis unit, collects and enters patient-related data to build a patient information data set, and analyzes and obtains the diversity of patient information in the patient information data set , according to diversity Matching a corresponding data encryption scheme from a data encryption scheme library to encrypt the stored patient information;
[0122] The patient classification unit classifies patients into different patient categories based on the data in the patient information data set, and uses the trained decision tree algorithm to identify key factors affecting the health of patients based on the patient-related data and disease characteristics, and outputs the identified key factors;
[0123] Health status assessment unit, which generates body status coefficients from patient health index data , according to the health status coefficient Mark the corresponding patients as unhealthy patients or sub-healthy patients, set the health range for sub-healthy patients, and use the health indicators and corresponding key factors as input to match the corresponding rehabilitation programs for unhealthy patients;
[0124] The prediction optimization unit collects the patient's health indicator data and then makes a prediction. If there are abnormal indicators in the obtained prediction values, the clustering of the abnormal indicators is generated according to the distribution state of the abnormal indicators. The patients are re-differentiated according to the clustering, and the rehabilitation plan for unhealthy patients is optimized.
[0125] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.
[0126] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0127] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0128] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only for some logical functions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0129] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0130] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0131] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.
[0132] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A method for managing chronic disease patient information based on a neural convolution algorithm, characterized in that: include, After collecting patient information management requirements, the complexity of the requirements is generated from the information management requirements data set If the complexity of the requirement If the complexity threshold is exceeded, a data collection instruction is issued to the outside; among them, the communication frequency and data update frequency Perform linear normalization and generate the required complexity as follows : ; Weight coefficient, , ; After collecting and entering patient-related data, construct a patient information data set, and analyze and obtain the diversity of patient information in the patient information data set , according to diversity Matching a corresponding data encryption scheme from a data encryption scheme library to encrypt the stored patient information; Differentiate patients into different patient categories based on the data in the patient information data set, and use the trained decision tree algorithm to identify key factors affecting patient health based on patient-related data and symptom characteristics; Generate health status coefficients from patient health indicator data , according to the health status coefficient Mark the corresponding patients as unhealthy patients or sub-healthy patients, set the health range for sub-healthy patients, and use the health indicators and corresponding key factors as input to match the corresponding rehabilitation programs for unhealthy patients; After collecting the patient's health indicator data, a prediction is made. If there are abnormal indicators in the obtained predicted values, the aggregation degree is generated according to the distribution state of the abnormal indicators. , according to the concentration Re-classify patients and optimize rehabilitation programs for non-healthy patients.
2. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 1, characterized in that: The frequency of communication between doctors and patients, the frequency of updating of patients' health data, and the change status data of patients' health status are counted within the preset patient information management cycle, and the information management demand data set is summarized and generated. The demand complexity is generated from the patient's information management demand data set .
3. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 2, characterized in that: After receiving the data collection instruction, collect the patient's basic information, medical record form, examination result form and follow-up record form, build a patient information data set, and set data access permissions after storing the patient information data set; Obtain several data encryption schemes with different cracking difficulties and build a data encryption scheme library in advance.
4. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 3, characterized in that: Calculate the diversity of patient information in a patient information dataset , where the similarity between the two patient information Perform linear normalization and map the corresponding data values to the interval In the following way: ; in, is the number of patients, It is The similarity between the patient information of the jth patient and the patient information of the jth patient, is the average similarity; weight coefficient: , ,and .
5. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 4, characterized in that: Preprocess the data in the patient information data set to obtain the preprocessed patient information to label the patients; use feature engineering to extract features from the patient information in each patient class to obtain the corresponding disease features; Collect relevant data for each patient class.
6. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 5, characterized in that: Collect and obtain the patient's health indicator data in the current stage and generate the patient's health indicator set; Generate the patient's health status coefficient from the health indicator set , if the patient's health status coefficient If the patient's blood pressure is lower than the health threshold, the corresponding patient is marked as an unhealthy patient, otherwise, he / she is marked as a sub-healthy patient.
7. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 6, characterized in that: Blood glucose concentration and blood pressure Perform linear normalization and generate the health status coefficient as follows : ; Weight coefficient: , ; is the number of collection points, is the blood pressure at the i-th acquisition node, is the mean blood pressure, is its qualified standard value; is the blood glucose concentration at the i-th collection node, is the mean blood glucose concentration, is the qualified standard value.
8. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 7, characterized in that: Taking several sets of current health indicator data as input, the trained neural convolutional network is used to predict the health trend of sub-healthy patients and obtain the corresponding prediction value; Set the healthy range for sub-healthy patients based on the obtained prediction values , if the predicted value of the health status coefficient of the sub-healthy patient is not within the healthy range Within, a prompt instruction is issued to the outside.
9. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 8, characterized in that: The acquired rehabilitation plans are summarized to build a rehabilitation plan library. The health indicators and corresponding key factors of non-healthy patients are used as input. The collaborative filtering algorithm is used to match the corresponding rehabilitation plans for non-healthy patients from the rehabilitation plan library, and the rehabilitation plans are sent to the non-healthy patients.
10. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 9, characterized in that: After the preset vital sign data collection cycle, the wearable device is used to monitor the patient's health indicator data in real time and summarize the patient's health data set; Use the time series prediction model to predict the data of various health indicators and obtain the corresponding indicator prediction values; use the indicator prediction values as input, and use the trained abnormal indicator recognition model to identify the abnormal indicators, output the identified abnormal indicators and determine the time node when the abnormal indicators occurred.
11. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 10, characterized in that: Combined with the distribution status of abnormal indicators in the current vital sign data collection period, the aggregation degree of abnormal indicators is generated according to the following method: : ; Weight coefficient: , ,and ; m is the number of abnormal indicators, It is i The abnormal indicators and j The time interval between abnormal indicators, is the average value of the time interval, is the acceptable value for the time interval; In the concentration When the concentration threshold is exceeded, if the patient is a sub-healthy patient, the sub-healthy patient is marked as an unhealthy patient. If the patient is an unhealthy patient, a program optimization instruction is issued to the outside.
12. The method for managing chronic disease patient information based on a neural convolution algorithm according to claim 11, characterized in that: After receiving the program optimization instruction, the change status data of the health indicators of non-healthy patients after executing the rehabilitation program and the feedback on the rehabilitation program are collected, and the reduction of the density of abnormal indicators is used as the optimization goal. The rehabilitation program of non-healthy patients is optimized using the trained multi-objective optimization algorithm to obtain the optimized rehabilitation program.
13. Chronic disease patient information management system based on neural convolution algorithm, characterized by: include, The demand analysis unit collects the patient information management requirements and generates the demand complexity from the patient information management requirements data set. If the demand complexity If the complexity threshold is exceeded, a data collection instruction is issued to the outside; Diversity analysis unit, collects and enters patient-related data to build a patient information data set, and analyzes and obtains the diversity of patient information in the patient information data set , according to diversity Matching a corresponding data encryption scheme from a data encryption scheme library to encrypt the stored patient information; The patient classification unit classifies patients into different patient categories based on the data in the patient information data set, and uses the trained decision tree algorithm to identify key factors affecting the health of patients based on the patient-related data and disease characteristics, and outputs the identified key factors; Health status assessment unit, which generates health status coefficients from patient health indicator data , according to the health status coefficient Mark the corresponding patients as unhealthy patients or sub-healthy patients, set the health range for sub-healthy patients, and use the health indicators and corresponding key factors as input to match the corresponding rehabilitation programs for unhealthy patients; The prediction optimization unit collects the patient's health indicator data and then makes a prediction. If there are abnormal indicators in the obtained prediction values, the clustering of the abnormal indicators is generated according to the distribution state of the abnormal indicators. The patients are re-differentiated according to the clustering, and the rehabilitation plan for unhealthy patients is optimized.
